ON MINIMAL NON-p-CLOSED GROUPS AND RELATED PROPERTIES
LUISE-CHARLOTTE KAPPE, MARCIN MAZUR, GABRIELA MENDOZA, AND MICHAEL B. WARD
Abstract. Let p be a prime. A group is called p-closed if it has a normal Sylow p-subgroup
and it is called p-exponent closed if the elements of order dividing p form a subgroup. A group
is minimal non-p-closed if it is not p-closed but its proper subgroups and homomorphic images
are. Similarly, a group is called minimal non-p-exponent closed if it is not p-exponent closed
but all its proper subgroups and homomorphic images are. In this paper we characterize finite
minimal non-p-closed groups and investigate the relationship between them and minimal
non-p-exponent closed groups. In particular, we show that every minimal non-p-closed
group is non-p-exponent closed and that minimal non-p-closed groups and simple minimal
non-p-exponent closed groups have cyclic Sylow p-subgroups. Furthermore, given a prime p,
we describe non-p-exponent closed groups of smallest order and we show that they coincide
with non-p-closed groups of smallest order.
Mathematics Subject Classification (2010). Primary 20D05, 20D25. Secondary 20F99.
Keywords. p-closed, p-exponent closed, non-p-closed, non-p-exponent closed, strongly pembedded.
1. Introduction
Let E be a group theoretic property. We say that a group G is a minimal non-Egroup if G is not an E-group but proper subgroups and proper homomorphic images of
G are E-groups. Already in 1903, Miller and Moreno [12] characterized finite minimal
non-E-groups for the property E = abelian. Their investigations were completed by
Rédei [15]. In [12] and [15], the same three authors considered the class of non-abelian
groups with every proper subgroup abelian. As it turned out, this class coincides with
the class of minimal non-abelian groups. Minimal non-E-groups have been investigated
for various properties such as nilpotency, solvability, and others. We refer to [6] for
further details.
1
2
LUISE-CHARLOTTE KAPPE, MARCIN MAZUR, GABRIELA MENDOZA, AND MICHAEL B. WARD
The topic of our investigation is the class of minimal non-E-groups, where E is either
the class of finite p-closed groups or the class of finite p-exponent closed groups. In [1],
Reinhold Baer defined the following property.
Definition 1.1. Let Σ be a set of primes. A group G is called Σ-closed if the set of
all elements of G whose order has no prime divisor outside of Σ is a subgroup of G.
In particular, a group is called p-closed if it is Σ-closed for Σ = {p}. Note that a finite
group is p-closed if and only if it has a normal Sylow p-subgroup. In particular, all
abelian groups are p-closed.
In [9], n-exponent closed groups are defined as follows.
Definition 1.2. Let G be a group and let n be a positive integer. Consider the set
G[n] = {g ∈ G : g n = 1} and let Gn = hG[n]i be the subgroup of G generated by G[n].
We say that G is n-exponent closed provided G[n] = Gn .
Clearly, an abelian group is n-exponent closed for any n. In particular, the concept of
an n-exponent closed group has been studied in the case when G is a p-group and n
is a power of p. In this case, it is customary to write Ω{i} (G) or Λi (G) for G[pi ] and
Ωi (G) for Gpi .
The classes of p-closed groups and p-exponent closed groups appear to have some
similarities, but neither is a subclass of the other. For example, any finite p-group is
p-closed but there is a wealth of finite p-groups which are not p-exponent closed. An
example of smallest order is the wreath product W = Cp ≀ Cp of a cyclic group of order
p with itself. The group W has order pp+1 , exponent p2 , and it is generated by elements
of order p. Thus any finite group having W as a normal Sylow p-subgroup is p-closed
but not p-exponent closed. On the other hand, if q is a prime divisor of 2p − 1, then
the group
2
T = T (p, q) = ha, b|aq = bp = 1, b−1 ab = a2 i
is not p-closed but it is p-exponent closed, since T [p] = hbp i is the center of T .
In light of the above remarks, it comes somewhat as a surprise that there is a strong
connection between minimal non-p-closed groups and minimal non-p-exponent closed
groups. As we will show, any minimal non-p-closed group and any minimal non-pexponent closed group is either solvable or simple (Theorem 3.1). In the solvable case,
ON MINIMAL NON-p-CLOSED GROUPS AND RELATED PROPERTIES
3
any minimal non-p-closed group is also minimal non-p-exponent closed, while a solvable
minimal non-p-exponent closed group is either minimal non-p-closed or it is a p-group
(Theorems 3.1 and 4.1). The minimal non-p-exponent closed p-groups are 2-generated,
have class at least p, exponent p2 and a cyclic center. A full classification of these
p-groups seems out of reach at the moment.
Looking at simple groups, we show that every simple minimal non-p-closed group is
also a minimal non-p-exponent closed group (Theorem 3.5). We do not know if the
converse always holds. Looking carefully at the techniques of [10], where the simple
minimal non-7-closed groups are classified, one can verify that the converse holds for
p ≤ 7 (section 3). The proof relies on the classification of finite simple groups and it
is crucial for the argument that p is small. Obtaining a similar classification for all
primes p will require substantially new ideas. However, we show that simple minimal
non-p-closed groups have cyclic Sylow p-subgroups with pairwise trivial intersection
(Theorem 3.5; the fact that the Sylow p-subgroups are cyclic was established independently in [17]). We also prove that simple minimal non-p-exponent closed groups
have cyclic Sylow p-subgroups, but whether they have pairwise trivial intersection remains open (Theorem 4.3). Incidently, only the results in this paragraph rely on the
classification of finite simple groups.
To introduce our final results, something should be said about what led us to the
investigation of these two classes of groups. The inspiration came from a paper by
Desmond MacHale [11]. In this paper 47 properties of groups are considered and
for each property MacHale asks for groups of smallest order which do not have the
property. The first property in [11] concerns groups in which the set of all squares of
elements is a subgroup. As can be easily seen, the alternating group A4 and the group
T (2, 3) mentioned above are exactly the groups of smallest order which do not have
this property. A natural extension of this result is to consider groups in which the set
of all n-th powers of elements is a subgroup. Such groups are called n-power closed
[9]. In [8], groups of smallest order which are not n-power closed have been determined
for any n not divisible by 16. A dual problem is to determine groups of smallest order
which are not n-exponent closed. To solve this problem in the case n is a prime was
the original motivation for our investigation. We show that G is a non-p-closed group
of smallest order if and only if it is a non-p-exponent closed group of smallest order and
4
LUISE-CHARLOTTE KAPPE, MARCIN MAZUR, GABRIELA MENDOZA, AND MICHAEL B. WARD
then G is either solvable or isomorphic to P SL(2, p) (Theorem 5.3). As a corollary of
our description of non-p-exponent closed groups of smallest order we get the following
formula for their order f (p):
1
f (2) = 6 and f (p) = min{p(kp + 1), p(p2 − 1)},
2
where p is an odd prime and k is the smallest positive integer such that kp + 1 is
a prime power (Corollary 5.4). This formula leads us to an old problem in analytic
number theory about the smallest prime in an arithmetic progression, which we discuss
in more detail in the last section.
2. A family of finite solvable groups
In this section we introduce a family of finite solvable groups which will play a major
role in the next section, where we characterize minimal non-p-closed groups.
Let p and q be distinct primes. Denote by ep (q) the order of q modulo p. Let F be
a finite field with q ep (q) elements. The multiplicative group F× of F is cyclic of order
q ep (q) − 1. Since p divides q ep (q) − 1, the group F× has a unique subgroup S of order p.
The group S acts on the additive group F of F by multiplication. Denote by U (p, q)
the semidirect product F ⋊ S. We have the following result.
Proposition 2.1. Let p and q be distinct primes and let F , S, and ep (q) be as defined
above. Then the following hold:
(i) The group S acts irreducibly on F ;
(ii) Let P be a cyclic group of order p. If A is an elementary abelian q-group
of order q m on which P acts irreducibly, then m = ep (q) and the semidirect
product A ⋊ P is isomorphic to U (p, q).
(iii) U (p, q) is minimal non-p-closed and minimal non-p-exponent closed.
Proof. Let Fq be the prime subfield of F. Let u be a generator of S. We claim
that Fq [u] = F. In fact, Fq [u] is a field with q s elements for some s ≤ ep (q). Thus
uq
s −1
= 1 and therefore p|q s − 1. It follows from the definition of ep (q) that ep (q)|s.
Consequently, s = ep (q) and F = Fq [u]. Any S-invariant subgroup of F is invariant
under multiplication by any element of Fq [u]. Since Fq [u] = F, it follows that {0} and
F are the only S-invariant subspaces of F . This shows that S acts irreducibly on F .
ON MINIMAL NON-p-CLOSED GROUPS AND RELATED PROPERTIES
5
Suppose now that P is a cyclic group of order p and A is an elementary abelian
q-group of order q m on which P acts irreducibly. Let c be a generator of P . Then A is
a simple module over the polynomial ring Fq [x], where x acts on A as c. It follows that
A can be identified with Fq [x]/J for some maximal ideal J of Fq [x]. Since the ideal J
is maximal, the ring Fq [x]/J is a field which we denote by F. Thus A is identified with
the additive group F of F. Moreover, the image of x in F is an element of order p and
the action of P on A is identified with the action of S on F by multiplication, where S
is the cyclic subgroup of F× of order p. Let u be a generator of S. The additive group
of the subfield Fq [u] of F is an S-invariant subgroup of F . Since F is irreducible, we
m
have Fq [u] = F. It follows that m is the smallest positive integer such that uq = u, i.e.
uq
m −1
= 1. Equivalently, m is the smallest positive integer such that p divides q m − 1,
i.e. m = ep (q). This proves that A ⋊ P is isomorphic to U (p, q).
Since Sylow p-subgroups of U (p, q) are cyclic of order p and not normal, the group
U (p, q) does not have any non-trivial normal p-subgroups. Hence it is neither p-closed
nor p-exponent closed. It follows from (i) that every proper subgroup of U (p, q) is
abelian and F is the only non-trivial proper normal subgroup. Now (iii) follows easily.
¤
3. Minimal non-p-closed groups
In this section we characterize minimal non-p-closed groups. Recall that if E is a
property of finite groups then a minimal non-E-group is any finite group G which does
not have the property E but the proper subgroups and proper quotients of G have
property E. For some properties E, e.g. abelian, if all proper subgroups of a group
have property E, then all proper quotients of this group also have property E. This
is not the case for the property of being p-closed. Consider the group T = T (p, q)
introduced in Section 1. All proper subgroups of T are abelian, hence are p-closed, but
the quotient of T by its center is not p-closed. Non-p-closed groups with all proper
subgroups p-closed have been called inner-p-closed (see e.g. [10]). Thus T (p, q) is
inner-p-closed but not a minimal non-p-closed group.
The following theorem describes minimal non-p-closed groups.
6
LUISE-CHARLOTTE KAPPE, MARCIN MAZUR, GABRIELA MENDOZA, AND MICHAEL B. WARD
Theorem 3.1. A finite group G is a minimal non-p-closed group if and only if it
satisfies one of the following conditions:
(i) G is a simple group which is a minimal non-p-closed group;
(ii) G is isomorphic to U (p, q) for some prime q different from p.
Proof. Let G be a minimal non-p-closed group which is not simple and let P be a
Sylow p-subgroup of G . Let A be a minimal normal subgroup of G, then A is a proper
subgroup and the quotient G/A is p-closed. Thus G/A has a normal Sylow p-subgroup.
It follows that all Sylow p-subgroups of G are contained in AP . Since G is not p-closed,
neither is AP . Thus AP = G, since all proper subgroups of G are p-closed. It follows
that G/A is a p-group.
Note that A is a direct product of copies of a simple group H. We claim that p
does not divide the order of H. In fact, suppose that p divides |H|. Then, since H is
simple and p-closed, H must be abelian. Thus A is an elementary abelian p-group. It
follows that G is a p-group, a contradiction. We conclude that |A| is prime to p. Thus
G = AP is a semidirect product. For any maximal subgroup M of P , the subgroup
AM of G is p-closed and M is a Sylow p-subgroup of AM . It follows that M is normal
in AM . Thus A normalizes M . Since maximal subgroups of P are normal in P , we
see that M is normal in G. Now if M is non-trivial, then G/M is p-closed. Since M is
a p-group, it follows that G is p-closed, a contradiction. Thus all maximal subgroups
of P must be trivial, hence P is cyclic of order p.
Suppose that H is non-abelian. For any prime divisor q of H, the group P acts by
conjugation on the set of Sylow q-subgroups of A. Since the number of such subgroups
divides |A|, it is prime to p and therefore the action has a fixed point. In other words,
P normalizes a Sylow q-subgroup Sq of A, which is also a Sylow q-subgroup of G.
Since Sq is a proper subgroup of A, the group Sq P is a proper subgroup of G, hence
it is p-closed. Thus P is normal in Sq P , i.e. Sq normalizes P . It follows that the
normalizer of P contains a Sylow q-subgroup of G for every prime q dividing |G|. This
however means that P is normal in G, a contradiction. Thus H must be abelian. It
follows that A is an elementary abelian q-group for some prime q different from p.
Being a minimal normal subgroup of G, the group A is an irreducible P -module. By
ON MINIMAL NON-p-CLOSED GROUPS AND RELATED PROPERTIES
7
Proposition 2.1, the group G is isomorphic to U (p, q). Conversely, the groups U (p, q)
are minimal non-p-closed groups by Proposition 2.1.
¤
The question arises if there are non-abelian simple groups which are minimal non-pclosed, and if there exist such groups, can they be classified? It follows from Dickson’s
Theorem (see e.g. [6, II.8.2]) that for any prime p > 3 the group P SL(2, p) is a simple
minimal non-p-closed group. For small primes p the classification of finite simple groups
can be employed to list all simple minimal non-p-closed groups. The main idea is a
trivial observation that simple minimal non-p-closed group cannot have proper simple
subgroups of order divisible by p. When p is small, the groups minimal among the
simple groups of order divisible by p are easy to list. Using known properties of these
groups one verifies which among them are minimal non-p-closed. The details have been
worked out for p ≤ 7. From the classification in [10] and the related results quoted
therein, we get the following description of simple minimal non-p-closed groups for
p ≤ 7:
(i) all minimal non-2-closed groups are solvable;
(ii) simple minimal non-3-closed groups are exactly the groups P SL(2, 2q ), q an
odd prime;
(iii) simple minimal non-5-closed groups are exactly the groups P SL(2, 5) and
Sz(2q ), q an odd prime;
(iv) simple minimal non-7-closed groups are exactly the groups P SL(2, 7), P SL(2, q)
for any prime q ≡ −1 (mod 7), and P SL(2, q 3 ) for any prime q ≡ 3 or 5 (mod 7).
A similar classification for all p > 7 seems out of reach in general. However, we show
that simple minimal non-p-closed groups have cyclic Sylow p-subgroups with pairwise
trivial intersection. The proof is based on the classification of simple groups with a
strongly p-embedded proper subgroup. Recall that a subgroup H of a group G is called
strongly p-embedded if for any non-trivial subgroup Q of some Sylow p-subgroup of G,
the normalizer of Q in G is contained in H.
Proposition 3.2. Let G be a simple minimal non-p-closed group.
(i) Any two distinct Sylow p-subgroups of G have trivial intersection.
(ii) The normalizer of a Sylow p-subgroup of G is a strongly p-embedded subgroup.
8
LUISE-CHARLOTTE KAPPE, MARCIN MAZUR, GABRIELA MENDOZA, AND MICHAEL B. WARD
Proof. Let P1 , P2 be two distinct Sylow p-subgroups of G such that the order of Q =
P1 ∩ P2 is largest possible. It suffices to show that Q is the trivial group. Suppose
otherwise and consider the normalizer N of Q in G. Since G is simple, N is a proper
subgroup of G. It follows that N has a normal Sylow p-subgroup S, which is contained
in a Sylow p-subgroup P of G. Now the normalizer Qi of Q in Pi is a p-group strictly
containing Q and contained in N , i = 1, 2. Thus Qi ⊆ S ⊆ P , hence P ∩ Pi has
order larger than the order of Q. This however implies that P = Pi and consequently
P1 = P2 , a contradiction. This completes the proof of (i).
Consider now a Sylow p-subgroup P of G and a non-trivial subgroup Q of P . If x ∈ G
normalizes Q, then Q is contained in both P and xP x−1 and therefore P = xP x−1 by
(i). In other words, the normalizer of Q is contained in the normalizer of P . Thus (ii)
holds.
¤
In [16], a classification of simple groups of Lie type with a strongly p-embedded
proper subgroup has been obtained (see, in particular, Theorem 7 therein). Using the
classification of finite simple groups and some facts about the sporadic simple groups
(contained, for example, in [3]), one can extend this to a list of all simple groups with a
strongly p-embedded proper subgroup (see [5, Th. 4.249]). Note that both in [16] and
[5] the classification is obtained for groups with simple generalized Fitting subgroup.
Restricting to simple groups, we obtain the following result.
Theorem 3.3. Let G be a finite simple group with a strongly p-embedded proper subgroup. Then one of the following conditions holds:
(i) G has cyclic Sylow p-subgroups;
(ii) G ∼
= P SL(2, q), P SU (3, q), Sz(q), 2 G2 (q), or A2p , and q is a power of p;
(iii) p = 3 and G ∼
= P SL(3, 4), or M11 ;
(iv) p = 5 and G ∼
= 2 F4 (2)′ , M c, or M (22);
(v) p = 11 and G ∼
= J4 .
Recall that a finite group is quasi-simple if it is perfect and its quotient modulo the
center is simple. Theorem 3.3 allows us to prove the following result.
ON MINIMAL NON-p-CLOSED GROUPS AND RELATED PROPERTIES
9
Corollary 3.4. Let p be an odd prime and let G be a finite simple group with a strongly
p-embedded proper subgroup such that the simple quotient of any proper quasi-simple
subgroup of G has order prime to p. Then G has cyclic Sylow p-subgroups.
Proof. The group G is one of the groups described in (i)-(v) of Theorem 3.3. The groups
in (iii)-(v) are eliminated by using the Atlas [3] to verify that they have proper nonabelian simple subgroups of order divisible by p. If G is one of the groups of (ii), then
it is isomorphic to P SL(2, p), P SU (3, p), or p = 3 and G ∼
= 2 G2 (27). For otherwise,
it clearly has a proper non-abelian simple subgroup of order divisible by p (note that
Sz(q) is eliminated since p 6= 2). Using the Atlas, we see that 2 G2 (27) has proper
non-abelian simple subgroups of order divisible by 3, hence it is not isomorphic to G.
For p ≥ 5, the group P SU (3, p) has a proper subgroup isomorphic to SL(2, p) ([13]).
From the Atlas we see that the group P SU (3, 3) has a proper subgroup isomorphic to
P SL(2, 7). It follows that G can not be isomorphic to P SU (3, p). This shows that if
G is one of the groups of (ii), then G ∼
= P SL(2, p). Since P SL(2, p) has cyclic Sylow
p-subgroups, the result follows.
¤
We now obtain the following result.
Theorem 3.5. Let G be a minimal non-p-closed group. Then G has cyclic Sylow
p-subgroups with pairwise trivial intersections. In particular, G is minimal non-pexponent closed.
Proof. According to Theorem 3.1, G is either simple or isomorphic to U (p, q) for some
q. In the latter case, the first part of the theorem is clear. Suppose then that G is
simple. We claim that p is odd in this case. Indeed, suppose that p = 2 and let x
be an element of order 2. By the Baer-Suzuki Theorem [4, Theorem 3.8.2], there is a
conjugate y of x in G such that the group hx, yi is not a 2-group. Thus hx, yi is not
2-closed and being a dihedral group, it must be a proper subgroup of G. This shows
that G is not a minimal non-2-closed group. Thus p is odd and the first part of the
theorem follows from Proposition 3.2 and Corollary 3.4.
To justify the last part, note that neither a simple group of order divisible by p nor
U (p, q) is p-exponent closed. Since G is minimal non-p-closed, any proper subgroup H
of G has a cyclic, normal Sylow p-subgroup. Hence H is p-exponent closed.
¤
10
LUISE-CHARLOTTE KAPPE, MARCIN MAZUR, GABRIELA MENDOZA, AND MICHAEL B. WARD
4. Minimal non-p-exponent closed groups
In this section we characterize minimal non-p-exponent closed groups. It turns out
that these groups are closely related to the minimal non-p-closed groups investigated in
the preceding section. We point out however that, unlike the class of p-closed groups,
the class of p-exponent closed groups is not quotient closed, as can be seen from the
groups T (p, q) introduced in Section 1. The main theorem of this section is very similar
to Theorem 3.1 except that the list of minimal non-p-exponent closed groups includes
some p-groups.
Theorem 4.1. A finite group G is a minimal non-p-exponent closed group if and only
if it satisfies one of the following conditions:
(i) G is a simple group which is a minimal non-p-exponent closed group;
(ii) G is a p-group which is a minimal non-p-exponent closed group. In this case G
is 2-generated, has class at least p, exponent p2 , and cyclic center. In particular,
|G| ≥ pp+1 ;
(iii) G is isomorphic to U (p, q) for some prime q 6= p.
Proof. Let G be a minimal non-p-exponent closed group which is not simple. There
are two elements x, y ∈ G of order p such that the order of xy is neither 1 nor p. The
minimality of G ensures that G is generated by x and y. Let A be a minimal normal
subgroup of G. Since G is not simple, the subgroup A is a non-trivial proper subgroup.
Let G = G/A. Again by minimality of G, the group G is p-exponent closed. Since G is
p-exponent closed, the set G[p] is a subgroup of G. Note that both xA and yA belong
to G[p]. Since G is generated by x and y, we conclude that G = G[p]. In other words,
G/A is a group of exponent p.
Note that A is a direct product of copies of a simple group H. Suppose that p divides
|H|. The group H, being isomorphic to a proper subgroup of G, is a p-exponent closed
group. Since H is simple, it must be abelian. It follows that A is an elementary abelian
p-group and therefore G is a p-group. Consequently, the group A contains a non-trivial
element c of order p which is central in G. Since A is a minimal normal subgroup of
G, we have A = hci. Thus A is cyclic of order p and it is central in G. It follows
that G has exponent p2 and therefore xy has order p2 . Since G/A has exponent p, we
ON MINIMAL NON-p-CLOSED GROUPS AND RELATED PROPERTIES
11
have (xy)p is a nontrivial element of A, hence A = h(xy)p i. This together with our
assumption that A is an arbitrary minimal normal subgroup of G shows that A is the
unique minimal normal subgroup of G. Consequently, A coincides with the center of
G. Since p-groups of class less than p are regular and regular p-groups are p-exponent
closed, the class of G is at least p. Clearly a group of class p has order at least pp+1 .
This establishes (ii) of the theorem.
Suppose now that p does not divide the order of H. Then |A| is prime to p and
G/A is a p-group. It follows that G is isomorphic to a semidirect product A ⋊ G/A. In
particular, the Sylow p-subgroups of G have exponent p. Note that for groups with a
Sylow p-subgroup of exponent p, to be p-closed is the same as to be p-exponent closed.
It follows that G is a minimal non p-closed group. By Theorem 3.1, the group G is
isomorphic to a group of the form U (p, q). Conversely, the groups U (p, q) are minimal
non-p-exponent closed by Proposition 2.1.
¤
Two questions naturally arise: what are the minimal non-p-exponent closed p-groups
and what are the simple minimal non-p-exponent closed groups? For the first question,
consider the wreath product W = Cp ≀ Cp , where Cp is a cyclic group of order p. The
group W is a minimal non-p-exponent closed group of order pp+1 . Indeed, this group
has exponent p2 and is generated by elements of order p (see [6, III.10.2]). Every
proper subgroup or quotient group of W has order at most pp hence it is a p-exponent
closed group. We do not know of any other construction of minimal non-p-exponent
closed p-groups. Note that Theorem 4.1(ii) implies that for each p there is only a finite
number of p-groups which are minimal non-p-exponent closed. It would be nice to have
an explicit description of all such groups.
As for the second question, it follows from Dickson’s Theorem (see e.g. [6, II.8.2])
that for any prime p > 3, the group P SL(2, p) is a simple minimal non-p-exponent
closed group. By Theorem 3.5, every simple minimal non-p-closed group is also a
minimal non-p-exponent closed group. We do not know if the converse holds for all
p. Using the methods of [10], one can verify that this is indeed true for p ≤ 7. In
general, we will show that simple minimal non-p-exponent closed groups have cyclic
Sylow p-subgroups, but we have not been able to show that different such subgroups
have trivial intersection. The key is the following observation.
12
LUISE-CHARLOTTE KAPPE, MARCIN MAZUR, GABRIELA MENDOZA, AND MICHAEL B. WARD
Proposition 4.2. Let G be a simple minimal non-p-exponent closed group. Then the
following hold:
(i) If P1 , P2 are different Sylow p-subgroups of G, then either they have trivial
intersection or Ω1 (P1 ) = Ω1 (P2 );
(ii) If P is a Sylow p-subgroup of G, then the normalizer of Ω1 (P ) is a strongly
p-embedded subgroup.
Proof. Let P1 , P2 be two different Sylow p-subgroups of G such that Ω1 (P1 ) 6= Ω1 (P2 )
and the group Q = Ω1 (P1 ) ∩ Ω1 (P2 ) has largest possible order. It suffices to show that
Q is the trivial group. Suppose otherwise and consider the normalizer N of Q in G.
Since G is simple, N is a proper subgroup of G. The normalizer Qi of Q in Ω1 (Pi )
contains Q properly and is contained in N , i = 1, 2. Let B = N [p]. Thus B is a
normal p-subgroup of N and it contains Qi , i = 1, 2. Let P be a Sylow p-subgroup of
G containing B. Then B ⊆ Ω1 (P ). It follows that the order of Ω1 (P ) ∩ Ω1 (Pi ) is larger
than the order of Q. This however implies that Ω1 (P ) = Ω1 (Pi ) and consequently
Ω1 (P1 ) = Ω1 (P2 ), a contradiction. This completes our proof of (i).
Consider now a Sylow p-subgroup P of G and a non-trivial subgroup Q of P . If
x ∈ G normalizes Q, then Q is contained in both P and xP x−1 and therefore Ω1 (P ) =
Ω1 (xP x−1 ) = xΩ1 (P )x−1 by (i). In other words, the normalizer of Q is contained in
the normalizer of P .
¤
The last proposition combined with Corollary 3.4 yields now the following result.
The proof is similar to the proof of Theorem 3.5.
Theorem 4.3. Let G be a simple minimal non-p-exponent closed group. Then G has
cyclic Sylow p-subgroups.
5. Non-p-exponent closed groups of smallest order.
We have not been able to characterize all simple groups which are minimal non-pclosed or minimal non-p-exponent closed groups. We can however describe such groups
of smallest possible order. The key is the following theorem, conjectured by E. Artin
and proved by R. Brauer and W. F. Reynolds [2].
ON MINIMAL NON-p-CLOSED GROUPS AND RELATED PROPERTIES
13
Theorem 5.1. Let G be a non-abelian simple group of order |G| divisible by a prime p
such that |G| < p3 . Then p > 3 and either G is isomorphic to P SL(2, p) or p = 2n + 1
is a Fermat prime and G is isomorphic to P SL(2, 2n ).
Corollary 5.2. Let p be a prime and let G be a finite non-abelian simple group of
smallest possible order divisible by p. If p > 3, then G is isomorphic to P SL(2, p).
In particular, G is both a minimal non-p-closed group and a minimal non-p-exponent
closed group. If p ∈ {2, 3}, then G is isomorphic to the alternating group A5 .
Proof. The conclusion is clear for p ∈ {2, 3}. Note that for p ∈ {2, 3}, the group A5 is
neither a minimal non-p-closed group nor a minimal non-p-exponent closed group.
Suppose that p > 3. Since p divides |P SL(2, p)|, we have |G| ≤ |P SL(2, p)| < p3
and Theorem 5.1 applies. If p = 5, then G ∼
= P SL(2, 5) ∼
= P SL(2, 22 ). If p > 5 and
p = 2n + 1 is a Fermat prime, then
1
|P SL(2, 2n )| = (22n − 1)2n = p(p − 2)(p − 1) > p(p2 − 1) = |P SL(2, p)|.
2
The first part of the corollary is now an immediate consequence of Theorem 5.1. The
second part follows immediately from Dickson’s Theorem [6, II.8.27].
¤
For a prime p, let n(p) = q m be the smallest prime power congruent to 1 modulo p,
where q = q(p) is a prime. Clearly, n(p) = kp+1, where k is the smallest positive integer
such that kp + 1 is a power of a prime. The following theorem describes non-p-closed
groups of smallest order and non-p-exponent closed groups of smallest order.
Theorem 5.3. Let p be a prime. The following conditions for a finite group G are
equivalent:
(i) G is a non-p-closed group of smallest order;
(ii) G is a non-p-exponent closed group of smallest order;
(iii) Either
1 2
(p
2
m
n(p) = q
− 1) ≤ n(p) and G is isomorphic to P SL(2, p) or
1 2
(p
2
− 1) >
and G is isomorphic to U (p, q).
Proof. Since |P SL(2, p)| < pp+1 , a non-p-exponent closed group of smallest order is
not a p-group by Theorem 4.1(ii). Note that P SL(2, 2) ∼
= U (2, 3) ∼
= S3 , P SL(2, 3) ∼
=
U (3, 2) ∼
= A4 , and
1 2
(p
2
− 1) 6= n(p) for p > 3. These observations together with
Theorems 3.1, 4.1, and Corollary 5.2 easily imply the theorem.
¤
14
LUISE-CHARLOTTE KAPPE, MARCIN MAZUR, GABRIELA MENDOZA, AND MICHAEL B. WARD
As a corollary to Theorem 5.3 we get an answer to the question which led us into
the investigations of this paper: given a prime p, what is the smallest order of a nonp-exponent closed group?
Corollary 5.4. Let p be a prime and let f (p) denote the smallest order of a non-pexponent closed group. Then

6,
f (p) =
min{p · n(p), 1 p(p2 − 1)},
2
if p = 2;
if p is odd.
Given a prime p > 3, it is natural to ask which of the groups P SL(2, p), U (p, q(p))
has smaller order. In other words, we would like to know which of the numbers n(p)
and 21 (p2 − 1) is smaller. With the help of a computer we checked that among the first
2000 primes only p = 19 satisfies n(p) > 12 (p2 − 1). It seems reasonable to conjecture
that 19 is the only such prime. This leads us to the following question: given an odd
prime p different from 19, is there a prime power which is congruent to 1 modulo p and
does not exceed 12 (p2 − 1)? This question is closely related to an old problem in number
theory which asks for the smallest prime in an arithmetic progression. It follows from
[14] that n(p) < 21 (p2 − 1) for infinitely many primes p. In fact, it is a long standing
conjecture in analytic number theory that if a and q are relatively prime and ǫ > 0
then the smallest prime p(q, a) congruent to a modulo q satisfies p(q, a) = O(q 1+ǫ ) (see
[7, Chapter 18]). If true, this conjecture implies that the equality f (p) = 21 p(p2 − 1)
can hold for only a finite number of primes p.
One can ask how often n(p) is a prime power rather than a prime itself. Clearly,
if p is a Mersenne prime, then n(p) = p + 1 is a power of 2. One can speculate
that for any prime q there are infinitely many primes p such that n(p) is a power of
q. To prove or disprove any such claim seems beyond the reach of current methods.
Perhaps more tractable is the question whether n(p) is not a prime for infinitely many
values of p. From the numerical evidence it seems that this happens rarely. For
example, among the first 2000 primes only the Mersenne primes 3, 7, 31, 127, 8191, and
the primes 13, 1093, 2801 yield n(p) which is not a prime.
We end this discussion with the following curious characterization of Mersenne primes
by group theoretic means.
ON MINIMAL NON-p-CLOSED GROUPS AND RELATED PROPERTIES
15
Corollary 5.5. For a prime p the following conditions are equivalent:
(i) p = 2 or a Mersenne prime;
(ii) There is a non-p-closed group of order p(p + 1).
Proof. Suppose that p = 2 or p is a Mersenne prime. Thus n(p) = p+1. By Theorem 5.3
and Corollary 5.4, the non-p-closed group of smallest order has order p(p + 1). Thus
(ii) follows from (i).
Conversely, suppose there is a non-p-closed group of order p(p + 1). Since p = 3 is
a Mersenne prime, we may assume that p > 3. Thus (p + 1) < (p2 − 1)/2 and we
conclude by Theorem 5.3 that n(p) ≤ p + 1. This is only possible if n(p) = p + 1 and
p is a Mersenne prime.
¤
Acknowledgements. The authors would like to thank Keith M. Jones for performing
computer computations of the numbers f (p) for the first 2000 primes p.
References
[1] R. Baer, Classes of finite groups and their properties, Illinois J. of Math. 1 (1957), 115-187.
[2] R. Brauer, W.F. Reynolds, On a problem of E. Artin, Ann. of Math. 68 (1958), 712-720.
[3] J.H. Conway, R.T. Curtis, S.P. Norton, R.A. Parker, R.A. Wilson, “Atlas of Finite Groups”, Clarendon
Press, Oxford, 1985.
[4] D. Gorenstein, Finite Groups, 2nd ed., Chelsea, New York, 1980.
[5] D. Gorenstein, “Finite Simple Groups: An Introduction to their Classification”, Plenum Press, New York,
1982.
[6] B. Huppert, “Endliche Gruppen I”, Die Grundlehren der Mathematischen Wissenschaften, Band 134,
Springer-Verlag, Berlin-New York 1967.
[7] H. Iwaniec, E. Kowalski, “Analytic Number Theory”, AMS Colloquium Pub. , 53, Providence, RI, 2004.
[8] L.-C. Kappe, G. Mendoza, Groups of minimal oreder which are not n-power closed, Contemp. Math. 511
(2010), 93-107.
[9] L.-C. Kappe, J.H. Ying, On exact power margin groups, Rend. Math. Univ. Padova 87 (1992), 245-265.
[10] X. Li, T. You, On the structure of inner 7-closed simple groups, J. of Math. Research & Exposition 17
(1997), 171-175.
[11] D. MacHale, Minimum counterexamples in group theory, Math. Mag. 54 (1981), 23-28.
[12] G.A. Miller, H. Moreno, Nonabelian groups in which every subgroup is abelian, Trans. Amer. Math. Soc.
4 (1903), 398-404.
16
LUISE-CHARLOTTE KAPPE, MARCIN MAZUR, GABRIELA MENDOZA, AND MICHAEL B. WARD
[13] H.H. Mitchell, Determination of the ordinary and modular ternary linear groups, Trans. Amer. Math.
Soc. 12 (1911), 207-242.
[14] Y. Motohashi, A note on the least prime in an arithmetic progression with a prime difference, Acta Arith.
17 (1970), 283-285.
[15] L. Rédei, Das schiefe Produkt in der Gruppentheorie, Comm. Math. Helvet. 20 (1947), 225-267.
[16] G.M. Seitz, Generation of finite groups of Lie type, Trans. Amer. Math. Soc. 271 (1982), 351-407.
[17] Xiao, Wen Jun, On p-closed groups, (Chinese) Acta Math. Sinica 36 (1993), no. 2, 160–162.
Department of Mathematical Sciences, Binghamton University, Binghamton, NY 13892-6000,
USA
E-mail address:
[email protected]
Department of Mathematical Sciences, Binghamton University, P.O. Box 6000, Binghamton,
NY 13892-6000, USA
E-mail address:
[email protected]
Department of Mathematics, Riverside Community College, Riverside, CA 92506-12999, USA
E-mail address:
[email protected]
Department of Mathematics, Western Oregon University, Monmouth, OR 97361, USA
E-mail address:
[email protected]